performance-based seismic retroﬁt strategy for existing

Performance-Based Seismic Retrofit Strategy for ExistingReinforced Concrete Frame Systems Using Fiber-Reinforced

Polymer CompositesStefano Pampanin1; Davide Bolognini2; and Alberto Pavese2

Abstract: The feasibility and efficiency of a seismic retrofit intervention using externally bonded fiber-reinforced polymer composites onexisting reinforced concrete frame systems, designed prior to the introduction of modern standard seismic design code provisions in themid-1970s, are herein presented, based on analytical and experimental investigations on beam-column joint subassemblies and framesystems. A multilevel retrofit strategy, following hierarchy of strength considerations, is adopted to achieve the desired performance. Theexpected sequence of events is visualized through capacity-demand curves within M-N performance domains. An analytical procedureable to predict the enhanced nonlinear behavior of the panel zone region, due to the application of CFRP laminates, in terms of shearstrength �principal stresses� versus shear deformation, has been developed and is herein proposed as a fundamental step for the definitionof a proper retrofit solution. The experimental results from quasistatic tests on beam-column subassemblies, either interior and exterior,and on three-storey three-bay frame systems in their as-built and CFRP retrofitted configurations, provided very satisfactory confirmationof the viability and reliability of the adopted retrofit solution as well as of the proposed analytical procedure to predict the actual sequenceof events.

DOI: XXXX

CE Database subject headings: Concrete, reinforced; Frames; Beam columns; Fiber reinforced polymers; Retrofitting.

Introduction

Extensive experimental-analytical investigations on the seismicperformance of existing reinforced concrete �RC� frame build-ings, primarily designed for gravity loads, as typically found inmost seismic-prone countries before the introduction of adequateseismic design code provisions in the 1970s, have confirmed theexpected inherent weaknesses of these systems �Aycardi et al.1994; Beres et al. 1996; Hakuto et al. 2000; Park 2002; Pampaninet al. 2002; Bing et al. 2002; Calvi et al. 2002a,b�. As a conse-quence of poor reinforcement detailing, lack of transverse rein-forcement in the joint region as well as absence of any capacitydesign principles, brittle failure mechanisms are expected. At alocal level, most of the damage is likely to occur in the beam-column joint panel zone while the formation of soft-story mecha-nisms can greatly impair the global structural performance ofthese RC frame systems. An appropriate retrofit strategy is there-fore required, which is capable of providing adequate protectionto the joint region while modifying the hierarchy of strengthsbetween the different components of the beam-column connec-tions according to a capacity design philosophy.

Alternative retrofit and strengthening solutions for reinforcedconcrete buildings have been studied in the past and adopted inpractical applications. A comprehensive overview of traditionalseismic rehabilitation techniques was presented by Sugano�1996�. Conventional techniques which utilize braces, jacketing,or infills as well as more recent approaches including base isola-tion and supplemental damping devices have been considered.Most of these retrofit techniques have evolved in viable upgrades.However, issues of costs, invasiveness, and practical implemen-tation still remain the most challenging aspects of these solutions.The results of numerical and experimental investigations on anoninvasive and economical retrofit solution based on metallichaunch connections have, for example, been recently presentedby Pampanin and Christopoulos �2003� and Pampanin et al.�2006�.

In the past decade, an increased interest in the use of advancednonmetallic materials, including shape memory alloys �Dolceet al. 2000� or fiber reinforced polymers �FRP� �FIB 2001, 2006�,has been observed. In this contribution, the feasibility and effi-ciency of a retrofitting intervention using FRP composite materi-als, according to a multilevel performance-based approach, willbe presented. Depending on the joint typology �interior or exte-rior� and on the structural details adopted, alternative objectives,in terms of hierarchy of strength and sequence of events withinthe beam-column-joint system, can be targeted and achieved.

After a summary of the experimental campaign on seismicvulnerability of existing underdesigned beam column subassem-blies and frame systems, representing the basic as-built con-figuration �benchmark� for this study and presented in previouspublications �Pampanin et al. 2002; Calvi et al. 2002a,b�, themain focus will be given herein to the description of �a� the prin-ciples and theoretical developments of the conceptual retrofitstrategy, �b� the main features of a simplified analytical model

1Dept. of Civil Engineering, Univ. of Canterbury, Christchurch, NewZealand. E-mail: [email protected]

2Dept. of Structural Mechanics, Univ. of Pavia, Italy.Note. Discussion open until September 1, 2007. Separate discussions

must be submitted for individual papers. To extend the closing date byone month, a written request must be filed with the ASCE ManagingEditor. The manuscript for this paper was submitted for review and pos-sible publication on September 21, 2005; approved on February 6, 2006.This paper is part of the Journal of Composites for Construction, Vol.11, No. 2, April 1, 2007. ©ASCE, ISSN 1090-0268/2007/2-1–XXXX/$25.00.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

181920

21

22232425262728293031323334353637383940

41

4243444546474849505152535455565758596061626364656667686970717273

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MARCH/APRIL 2007 / 1

adopted to evaluate the increase in the joint shear strength due tothe application of FRP, �c� the assessment of the internal hierar-chy of strength through M-N �moment-axial load� performancedomains to account for the variation of axial load in the column,�d� the implementation of the proposed solution, and �e� the ex-perimental validation of the intervention via quasistatic cyclictests on four beam-column joint subassemblies and one three-storey three-bay frame system, 2/3 scaled, retrofitted with CFRPsheets. Comparisons with the response of the benchmark �i.e.,as-built configuration� subassemblies and frame specimen are car-ried out to emphasize the enhanced behavior of the retrofittedconfigurations as well as the general reliability of the overallperformance-based seismic retrofit strategy.

Seismic Behavior of Existing Poorly DetailedRC Frames

Experimental Investigations on As-Built Systemsand Subassemblies

The first phase of the research project involved the assessment,through analytical and experimental investigations, of the seismicvulnerability of existing reinforced concrete frame systems, pri-marily designed for gravity loads as typically found in majorseismic prone countries in the period between the 1950s and the1970s, before the introduction of modern seismic design provi-sions in the mid-1970s.

In order to facilitate the introduction of the proposed retrofitstrategy as well as to provide a benchmark comparison for theexperimental tests on the FRP retrofitted configurations, presentedin the following paragraphs, a brief summary and overview of theexperimental results on the as-built solutions is given herein. Fur-ther details on the response of the as-built specimens can be foundin Pampanin et al. �2002� and Calvi et al. �2002b�, while moreextensive research reports and publications on the experimentaland numerical investigations on existing pre-1970 frames areunder preparation.

The experimental program on existing �as-built� RC framesubassemblies and systems comprised of quasistatic tests carriedout in the Laboratory of the Department of Structural Mechanicsof the University of Pavia on six, one-way beam-column jointsubassemblies �two exterior knee joints, two exterior tee joints,and two interior joints� as well as on a three-storey three-bay

frame system. Both beam-column subassemblies and frame sys-tems were scaled at 2 /3. Particular attention was given to thevulnerability of the panel zone region.

The design recommendations provided by the current Italiannational design provisions �Regio Decreto, 1939� were followedand, where necessary, integrated by textbooks broadly adoptedin the engineering practice and available in that period �e.g.,Santarella 1957�. It is worth noting that typical plan configura-tions of existing buildings in Mediterranean seismic-prone coun-tries would consist of frames running in one direction only withlightly reinforced slab �perforated clay brick units with cast-in-situ concrete topping� spanning in the orthogonal direction.Simple one-way beam-column joint subassembly specimenswithout transverse beam nor structural cast-in-situ slab are thusadequate representatives of such construction practice. Furtherexperimental investigation on the seismic performance of alterna-tive beam-column joint typologies including two-way exteriorbeam-column joints with and without cast-in-situ slab under bi-directional cyclic loading are currently ongoing at the Universityof Canterbury, Christchurch �Hertanto, 2006�.

Table 1 reports the geometric and reinforcement details of thecritical sections of the beam-column subassembly specimens. Inparticular, Figs. 1 and 2 show, respectively, the details of thebeam-column joint specimens T1 and C2, used as benchmarkconfigurations before the retrofit intervention, and of the twoframes �identical structural systems, tested in the as-built and ret-rofit configuration�.

Consistently with most of the old practice, no transverse re-inforcement was placed in the joint region. Plain round bars,with mechanical properties similar to those typically used in olderperiods, were adopted for both longitudinal and transverse re-inforcement. Beam bars in exterior joints were not bent into thejoint region, but anchored with endhooks. Lap splices with hookanchorages were adopted in the beam bars crossing the interiorjoints �except for the specimen C1 with continuum reinforcement�as well as in column longitudinal bars at each floor level abovethe joint region and at the column-to-foundation connection �inthe frame system�.

Test Setup and Loading Protocols

In order to allow for a comparison between the response of theas-built and the retrofitted configuration, some details in the test

Table 1. Specimen Reinforcement and Section Geometry

Jointtype Specimen

Sectiondimensions

�mm�Longitudinalreinforcement

Transversereinforcement

Exterior T1A,T1Bb Beam 330�200a Top 2�8+2�12; �4@115 mma

Bottom 2�8+2�12

Columna 200�200a 3�8+3�8a �4@13 mma

T2A T2Bb Beam Top 2�8+1�12;

Bottom 2�8+1�12

Interior C1b, C3b Beam Top 2�8+3�12;

Bottom 2�8+1�12

C2, C4 Beam Top 2�8+2�12;

Bottom 2�8+1�12aEqual reinforcement for all specimens.bReinforced with FRP.

74

757677787980818283848586

8788

8990

919293949596979899

100101102103104105106107108109110111112113

114

115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151

152

153154

2 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MARCH/APRIL 2007

setup and loading protocol were implemented for the tests on thebeam-column subassemblies and on the frame systems as dis-cussed in the following paragraphs,

The test setup �Fig. 3� and loading history/regime �Fig. 4� ofthe beam-column joint subassemblies were intended to accuratelyreproduce the actual configuration within a frame subjected toreversed cyclic lateral loading. Beam and column elements wereextended between points of contraflexure �assumed to be at mid-span in the beams and at midheight in the columns� where pinconnections were introduced. Simple supports at the beam endswere obtained connecting pin-end steel members to the floor.

The loading protocol consisted of a series of three cycles atincreasing levels of interstorey drift applied to the top of thecolumn through a horizontal hydraulic actuator. In order to moreclosely reproduce the actual stress level in the joint during thelateral cyclic sway of a frame building, the column axial load wasvaried during the experimental tests as a function of the lateralload, by means of a vertical hydraulic jack, acting on a steel plateconnected to the column base plate by vertical external post-tensioned bars.

The axial-load versus lateral-force relationships for exteriorand interior joints, which are functions of the geometric charac-teristics of the frame �i.e., bay number and length, number ofstoreys�, were evaluated with preliminary pushover analyses onthe three-storey–three-bay RC frame system. Significant varia-tions of the axial load up to 40–50% with respect to the value dueto gravity load only were expected. During the tests on the beam-column specimens, a simplified bilinear relationship betweenaxial and lateral load was adopted as shown in Fig. 4�b�. It isworth underlining that the adoption of a variable axial load rep-resents a fundamental, although unusual, improvement in theloading protocol typically adopted for quasistatic tests on existingbeam-column subassemblies available in literature, where theaxial load is most likely maintained constant. The importance of aproper estimation of the variation of the axial load, particularlywhen dealing with assessment and retrofit strategies of poorlydetailed RC frame subassemblies of system, will be more evidentafter the considerations given in the following sections, whendiscussing the delicate process of evaluating the hierarchy ofstrength and sequence of events.

The frame system was subjected to quasistatic cyclic loadingat increasing levels of floor displacements, applied to the structureusing three electromechanical actuators connected to the closestbeam through a steel extension arm. The presence of gravityloads, which in existing underdesigned or gravity-load dominatedframes represents a significant portion of the overall capacity, wassimulated using concrete blocks as shown in Fig. 5. The lateralloading history consisted of a series of three cycles at increasinglevel of top drift �±0.2%; ±0.6%; ±1.2%� with one conclusivecycle at ±1.6%. The application of simulated seismic loads wasbased on a hybrid force displacement control: the top floor dis-placement was directly controlled while maintaining a code-typelateral force distribution, proportional to the mass and to the floorlevel height �more details in Calvi et al. 2002b�.

Behavior of As-Built Beam-Column JointSubassemblies

As reported in Pampanin et al. �2002�, the exterior tee-joint speci-mens showed a particularly brittle failure mechanism given by thecombination of joint shear damage with the effects of slippage ofthe plain round beam longitudinal bars within the joint region,which led to a concentrated compressive force at the end-hookanchorage. As a result, a concrete “wedge” tended to spall off�Fig. 6�, leading to a brittle behavior with marked pinching in thehysteresis loop and loss of bearing-load capacity �Fig. 7, darkdashed line, top and center�.

Conversely, the interior joint specimens showed significant re-sources of plastic deformation �Fig. 7, dark dashed line, bottom�,even without specific ductile structural details. A marked pinchingwas still observed, due to slip of the column longitudinal re-inforcement bars. According to preliminary capacity designconsiderations, shear joint cracking and column hinging were pre-dicted to be relatively close events. The concentration of flexuraldamage in the column at early stages thus acted as a structuralfuse for the joint panel zone, which did not suffer significantcracking and damage. However, it should be recalled that theglobal frame system response could be seriously impaired if col-umn hinging led to a soft-storey mechanism.

Fig. 1. Geometry and reinforcement details in exterior joint specimen T1 and interior specimen C2 �as-built solutions�

155

156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194

195

196197198199200201202203204205206207208

209210

211212213214215216217218219220221222223224225226227228229230231


Global Behavior of the Frame System in the As-BuiltConfiguration

The results of the quasistatic tests on the three-storey–three-bayframe system �briefly summarized in Calvi et al. 2002b� con-firmed the high vulnerability of the panel zone region as observedat a subassembly level �particularly in exterior joints� and thetendency to develop undesirable global mechanisms, due to theabsence of an adequate hierarchy of strength.

As shown in Fig. 8, most of the damage concentrated in thejoint region �exterior tee-joints� or at the beam-column interfacesthrough the development of a single wide flexural crack as ex-pected from the slip of plain round reinforcing bars. In the interiorjoints no cracks were observed. The exterior tee-joints were sub-jected to a damage mechanism, analogous to that observed duringthe tests on beam-column subassemblies. The aforementionedtendency to develop a concrete wedge mechanism due to com-bined effects of an inefficient strut mechanism in the joint regionafter first shear cracking in the joint and the stress concentrationat the beam bar end hooks, was observed, which could lead tosevere damage and consequent loss of load-bearing capacity. The

test was, however, interrupted at relatively early stages �1.6%drift� for safety issues, after the clear indication of a softeningbehavior as shown by the hysteresis behavior in Fig. 9 �darkdashed line�.

At a global level, an interesting peculiar mechanism wasobserved, when compared to a weak-column strong-beam mecha-nism �which would lead to a soft storey mechanism�, typicallyexpected in an existing building. Based on the experimental evi-dence and numerical investigations, the concept of a shear hingemechanism has been proposed as an alternative to flexural plastichinging in the beams �Pampanin et al. 2002, 2003�. The concen-tration of shear deformation in the joint region, through the acti-vation of a so-called shear hinge, could in fact result beneficial,by spreading the interstorey drift demand among two consecutivestoreys, thus reducing the deformation demand onto the adjacentstructural members �columns in particular� and postponing theoccurrence of undesirable soft-storey mechanism. As noted byCalvi et al. �2002a�, the drawback of this apparently favorableeffect on the global response is, however, the increase in sheardeformations in the joint region which can possibly lead �depend-

Fig. 2. Test frame: geometric and mechanical characteristic

232

233

234235236237238239240241242243244245246247248249250251

252

253254255256257258259260261262263264265266267268269270271


ing on the joint typology and structural details adopted� tostrength degradation and loss of vertical load-bearing capacity.The post-cracking behavior of the joint depends, in fact, solely onthe efficiency of the compression strut mechanism to transfer theshear within the joint. Thus, while rapid joint strength degradationafter joint diagonal cracking is expected in exterior joints, a hard-ening behavior after first diagonal cracking can be provided by aninterior joint. Damage limit states based on joint shear deforma-tions have recently been defined and reported in Pampanin et al.�2003� as a support to seismic assessment and retrofit strategy ofpre-1970s RC frame systems. It is evident how, based on a de-tailed assessment of the local damage and corresponding globalmechanisms, a more reliable seismic rehabilitation strategy can bedefined.

Multilevel Retrofit Strategy

Regardless of the technical solution adopted, the efficiency of aretrofit strategy strongly depends on a proper assessment of theinternal hierarchy of strength of the beam-column joints as wellas of the expected sequence of events within a beam-column sys-

tem �shear hinges in the joints or plastic hinges in beam andcolumn elements�. The effects of the expected damage mecha-nisms on the local and the global response should also be ad-equately considered.

Performance-Based Retrofit Strategy

An ideal retrofit strategy would not only protect the joint panelzone region, by identifying �critically� weak point in older frames,but would further upgrade the structure to exhibit the desiredweak-beam strong-column behavior which is at the basis of thedesign of new seismic resistant RC frames. However, due to thedisproportionate flexural capacity, in gravity-load-dominatedframes, of the beams when compared to the columns, a completeinversion of hierarchy of strengths is difficult to achieve in allcases and for all beam-to-column connections without major in-terventions. This is more evident for interior beam-to-columnconnections where the moment imposed on interior columns fromthe two framing beams is significantly larger than for exteriorcolumns. As indicated in the previous paragraph, interior jointsare less vulnerable than exterior joints and exhibit a much morestable hysteretic behavior with hardening after first cracking.

Fig. 3. Test setup of exterior and interior beam-column joints

272

273274275276277278279280281282283284285

286

287288289290

291

292293294

295

296297298299300301302303304305306307308309310


It is thus conceivable, in a bid to protect the interior columnsfrom hinging, to tolerate some joint damage. According to amultilevel retrofit strategy approach suggested by Pampanin andChristopoulos �2003�, two levels of retrofits can therefore beconsidered, depending on whether or not the interior joints canbe fully upgraded. A complete retrofit would consist of a fullupgrade by protecting all joint panel zones and developingplastic hinges in beams while columns are protected according tocapacity design principles. A partial retrofit would consist ofprotecting exterior joints, forming plastic hinges in beams fram-ing into exterior columns, while permitting hinging in interiorcolumns or limited damage to interior joints, where a full reversalof the strength hierarchy is not possible. The viability of the par-tial retrofit strategy must be investigated on a case-by-case basisto assure that the localized damage to interior joints does notseverely degrade the overall response of the structure or jeopar-dize the ability of the interior columns to safely carry gravityloads.

Assessment of Sequence of Events:Performance Domains

A simple procedure to compare the internal hierarchy of strengthswithin a beam–column-joint system is herein presented. Theevaluation of the expected sequence of events is then proposed tobe carried out through comparison of capacity and demand curveswithin a M-N �moment-axial load� performance domain.

Fig. 4. Loading protocol for beam-column joint subassemblies:�a� top drift control; �b� axial load versus lateral load relationship

Fig. 5. Frame test setup and elevation view

Fig. 6. Evaluation of hierarchy of strengths and sequence ofevents: M-N performance domain �exterior tee-joint T1 in as-builtconfiguration�

311

312313314315316317318319320321322323324325326327328

329330

331332333334335


Fig. 6 shows, as an example, the M-N performance domainadopted to predict the sequence of events and level of damage inthe joint panel zone expected for the exterior specimen T1.

The capacities of beam, column, and joints are referred to agiven limit state �e.g., for joints: cracking, equivalent “yielding”or extensive damage, and collapse� and evaluated in terms ofthe equivalent moment occurring in the column at that stage,based on equilibrium considerations within the beam–column-joint specimen. While the evaluation of M-N curves for beamsand columns is a relatively simple task, the definition of an“equivalent” curve to represent the joint panel zone can rely onthe procedure described below.

The capacity or damage level of a joint is typically expressedin terms of nominal shear stress �� jn� or principal compression-tensile stresses �pc , pt�. Although current codes �e.g., ACI 318,AIJ, EC8, NZS3101� tend to adopt simplified provisions whichlimit the nominal shear stress � jn expressed as a function of theconcrete tensile strength, k1

�fc�, or the concrete compressivestrength, k2fc�, where k1 and k2 are empirical constants, it is com-monly recognized that principal stresses, by taking into accountthe contribution of the actual axial compression stress acting inthe column, can provide more accurate indications on the stressstate and thus damage level in the joint region.

Typical strength degradation models, available in the literatureand based on research on poorly designed joints �e.g.,Priestley1997, Pampanin et al. 2002; shown in Fig. 10� can be adopted todefine limit states in a joint panel zone subjected to shear andaxial load.

According to the simplified analytical model proposed byPampanin et al. �2003� to describe the joint nonlinear behavior,based on a rotational spring within a concentrated plasticity ap-

Fig. 7. Behavior of interior joint specimen C3: detachment of thecarbon fibers from the nodal region at 4% drift level

Fig. 8. Observed damage in as-built frame at 1.6% top drift

Fig. 9. Comparison of global hysteresis behavior �base shear topdrift� beween as-built and retrofitted frame

Fig. 10. Strength degradation curves for exterior joints in terms ofprincipal tensile stress versus shear deformation

336

337338339340341342343344345346347

348

349350351352

353354355356357358359360361362363364365366


proach, the equivalent moment-rotation curve of the joint region�i.e., monotonic characteristics of the spring model� can be de-rived from the corresponding principal tensile stress-shear defor-mation curve using equilibrium considerations: for any givenlevel of principal tensile �or compression� stress in the joint, thecorresponding “joint moment” Mj, which is either the sum ofthe beam moments or the sum of the column moments at thatstage, can be evaluated. So doing, M-N capacity curves cor-responding to the different joint limit states can be plotted withina performance domain where “equivalent column” capacity arerepresented.

As previously shown in Fig. 6 �as-built exterior specimen T1�,demand curves �V-shaped� should account for the variation ofaxial load due to the effects of lateral forces in a frame system�for either opening and closing of the joint�. Incorrect and non-conservative assessment of the sequence of events can otherwise

result, leading to inadequate design of the retrofit intervention.It is worth noting that, for simplicity, the sequence of eventscorresponding to negative and positive sign of the lateral force�opening or closing of the joint�, should better be independentlyevaluated �i.e., the numbering 1–8 actually indicates a sequence1–4 in the negative direction and a sequence 1�–4� in the positivedirection�.

In the case of specimen T1, in the as-built configuration, apure shear hinge mechanism, with extensive damage of the joint,was thus expected �using a proper demand curve� prior to anyhinging of beams or columns �Table 2�, as confirmed by the ex-perimental tests. However, the order and “distance” of the eventsstrongly depends on the demand curve assumed. If a constantaxial load curve was used �as shown in Fig. 6 for N=−100 kN�,only a minor increase in the column strength �in addition to thejoint strengthening� would have appeared necessary, leading to a

Table 2. Sequence of Events for Exterior Specimen T1 �As-BuiltConfiguration�

Specimen T1 �as-built�

Type oflateralforce No. Event

Lateralforce�kN�

1 Joint cracking and deterioration

starting pt=0.19�fe−

−10.94

Open jointF�0

2 Beam yielding −16.59

3 Upper column yielding −20.50

4 Lower column yielding −22.75

5 Joint failure 9.37

Close jointF�0

6 Lower column yielding 13.50

7 Upper column yielding 14.50

8 Beam yielding 16.59

Fig. 11. Effects of FRP on the moment-curvature curve of a member �beam� critical section

Fig. 12. Joint strength degradation curve: contributions of FRP andconcrete �exterior speciment T1�

367

368369370371372373374375376377378379380381382

383

384385386387388389390391392393394395396397398


column hinging occurring before the formation of a beam hinge�i.e., high risk of a soft storey mechanism even after the retrofitintervention�.

The concept of a performance domain could thus be extendedfrom the purpose of assessing as-built systems and adopted toevaluate and control the feasibility and efficiency of any retrofitstrategy on beam-column joints, provided that the effects of theretrofit solution on the single elements �beams, column, or jointpanel zone� can be simply and independently evaluated as de-scribed in the following paragraphs.

Evaluation of FRP Strengthening Effects:Analytical Model

The effects of a retrofit intervention with FRP composite materialin the form of externally bonded reinforcement on a beam-columnjoint, in terms of flexural or shear capacity in beams, columns,and panel zone region, is carried out through a step-by-step pro-cedure. The occurrence of defined limit states �cracking, yielding,debonding, crushing, and spalling of concrete, failure within theadopted materials� corresponding to a given stress or strain valuecan thus be properly evaluated and controlled when designing theretrofit intervention. As mentioned and shown, an accurate pre-diction of the expected sequence of events can thus be obtainedthrough M-N performance domains.

Analytical procedures available in the literature are adoptedand properly modified to account for debonding phenomena aswell as, more importantly, for the effects of the variation of axialload onto the joint panel zone behavior �critical issue typicallyneglected�.

Flexural FRP Retrofit of Beams and Columns

The enhanced flexural behavior of a FRP retrofitted beam or col-umn critical section was evaluated though a fiber section analysis.The Bernoulli-Navier hypothesis on plane sections remainingplane was assumed, considering fully composite action �bond�between the external FRP laminates and the concrete. Debondingwas taken into account according to the model proposed byHolzenkämpfer �1994� �and adopted by the FIB guidelines ofFRP retrofit, FIB 2001�, and thus expected to occur at a strainlimit level �deb=c1 ·�fctm /Eftf, where Ef is the FRP E-modulus,fctm the mean value of concrete tensile strength, s the thickness ofthe FRP laminate, and c1 an empirical coefficient taken as 0.64 forCFRP as suggested by Neubauer and Rostásy �1997�.

The material behavior was defined through proper stress-strainrelationships, as follows: Mander et al. �1988� model for concrete;Dodd-Restrepo model �1995� for steel and a linear-elastic rule forthe FRP composite material, consistent with the properties sup-plied by the provider.

The moment-curvature behavior of the critical section in thepresence of externally bonded FRP laminate can thus be evalu-ated for different levels of axial load �Fig. 11� using an iterativeprocedure as typically done for RC sections.

The position of the neutral axis is estimated until both com-patibility and equilibrium conditions are satisfied. M-N capacitycurves for beams and columns corresponding to a given limitstate can be derived and plotted in a performance domain to de-fine the sequence of events.

The confinement effects of the FRP on the section curvatureductility capacity can be taken into account following proceduresavailable in the literature �e.g., Spoelstra and Monti 1999�.

Table 3. Properties of High Modulus Carbon Fiber with Unidirectional Fabric �MBrace CFRP C5-30�

Type of fiberTow sheet

typeDensity�kg/m3�

Effectivethickness

�one layer��mm�

Tensilestrength�MPa�

E-modulus�MPa�

Utimatestrain�%�

High modulus carbon �C5-30� Unidirect. fabric 1820 0.165 3,000 39,000 0.8

Fig. 13. FRP-retrofit solution for the exterior joint specimen T1B

399

400401402403404405406407408

409410

411412413414415416417418419420421422423424425426

427

428429430431432433434435

436437438439440441442443444445446447448449450451452453454455456


Increase of Joint Shear Capacity due to the FRP

The evaluation through analytical models of the strengtheningeffects on the panel zone �joint� shear is a more complex taskwith limited research available in the literature. An overviewof alternative procedures has been given by Antonopoulos andTriantafillou �2002�. Typical �oversimplified� approaches considerthe contribution of the FRP equivalent to external “stirrups”�analogy with steel transverse reinforcement�. Upper limits of themaximum strain in the FRP material are used in the calculations,either corresponding to the declared ultimate tensile capacity�Gergely et al. 1998� or to constant strain values depending on thepreparation of the concrete surface �Tsonos and Stylianidis 1999;Gergely et al. 2000�.

A more rigorous model, based on stress equilibrium and straincompatibility equations of the panel zone region �idealized as athree-dimensional element� has been presented by Antonopoulosand Triantafillou �2002� as an extension of the model for RC jointbehavior without FRP proposed by Pantazopoulou and Bonacci�1994�. Satisfactory validation of the analytical model wasobtained on the experimental results on a total of 15 beam-column exterior beam-column subassemblies, tested by the au-thors �Antonopoulos and Triantafillou 2003� or available in theliterature �Gergely et al. 2000�.

It is, however, important to underline that, as typically done inmost experimental tests on beam-column joints, no variation ofaxial load as a function of the lateral force during the lateral swayof a frame system was considered during the tests. The implica-tions of assuming a constant load in the assessment of the se-quence of events prior to or after a retrofit intervention has beenbriefly discussed in the previous paragraphs.

In the present contribution the original step-by-step iterativeprocedure proposed by Antonopoulos and Triantafillou �2002�,in its simplified version �where the direct shear strength of thecomposite sheet is neglected�, is adopted as a general platformand adequately extended after a few simple modifications toaccount for the variation of the axial load on the joint region.Consistently with the analytical procedure proposed to visualizethe joint shear contribution within a M-N performance domainstarting from principal tensile or compression stresses consi-derations, the basic equations of equilibrium and strain compat-ibility of the joint panel zone are rearranged to evaluate an

equivalent strength degradation curve �principal tensile stressversus joint shear deformation� corresponding to the FRP contri-bution only.

The overall strength degradation curve for the FRP retrofittedjoint would thus be given by the combination of the FRP andconcrete contributions, as shown in Fig. 12. Such a curve formsthe basis for the evaluation of the equivalent joint moment Mj,within a performance domain M-N.

It is worth noting that in terms of analytical-numerical model-ing according to a plasticity-concentrated approach, two rota-tional springs �with moment-rotation curves derived, as men-tioned, according to the method proposed by Pampanin et al.�2003�� can be adopted to represent the two independent contri-butions.

It is in fact expected �later confirmed by the experimentaltests� that the cracking and damage of the joint can still occurunderneath the protection given by the FRP laminates, whosemajor effect is to increase the overall joint strength, avoidinglocal failure mechanism �such as the “concrete wedge” mecha-

Fig. 14. FRP-retrofit solution for the interior joint specimen C3

Fig. 15. Evaluation of hierarchy of strengths and sequence ofevents: M-N performance domain �exterior tee-joint T1B joint afterretrofit�

457

458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497

498

499500501502503504505506507508509510511512513514515516


nism� and achieving an enhanced global behavior by developing amore desirable sequence of events �e.g., weak-beam strong-column mechanism if a total retrofit strategy is followed�.

Details on the analytical procedure to evaluate the joint shearstrength contribution of FRP as well as on simplified designmethods can be found in Vecchietti �2001� and Nassi �2002�and will be reported in future publications currently underpreparation.

Design of the Retrofit Intervention

According to the proposed multilevel retrofit strategy, a full ret-rofit was adopted for the exterior joint, i.e., protection of the jointand plastic hinge in the beam, while a partial retrofit was adoptedfor the interior joint specimen, i.e., partial protection of the col-

umn hinging while some damage in the joint region can be ac-cepted. Issues related to the expulsion of the concrete wedge inthe exterior joints as well as to the premature debonding of thefibers were carefully considered as explained in the followingsections.

Retrofit Solutions

A few alternative FRP retrofit solutions �relying on differentforms or properties of the composite material� have been recentlyproposed in literature for beam-column joints subjected to lateralcyclic loading �e.g., Pantelides et al. 2000�. Extensive experimen-tal investigations on an exterior beam-column joint retrofittedwith FRP �in the form of laminates or strips� have been carriedout by Antonopoulos and Triantafillou �2003�. Due to the scope ofthat investigation �evaluation of the FRP contribution to the joint

Fig. 16. Preparation of FRP retrofit intervention on the test frame

Fig. 17. Comparison of damage mechanisms in beam-column joint subassemblies before and after retrofit; �a� shear hinge �as-built T1�;�b� relocated beam plastic hinge �retrofitted T1B�; �c� joint panel zone conditions underneath the composite sheet after testing

517

518519520521522523524

525

526527528529

530

531532533534

535

536537538539540541542543


shear strength�, the design of the retrofit strategy aimed at guar-anteeing that the damage occurred in the joint region. A selectiveseismic strengthening technique for gravity-load-designed frames,relying on both FRP laminates and near mounted surface, hasbeen recently proposed by Prota et al. �2002�.

In this contribution, unidirectional carbon fiber laminates�high-modulus CFRP, Table 3� were adopted for both exterior and

interior joints in the configurations illustrated in Figs. 13 and 14.It is worth noting that after considering alternative FRP materials�i.e., glass, aramid, or carbon fibers with lower modulus andstrength�, the choice of high modulus CFRP, with a relativelylow ultimate strain capacity, was primarily dictated by the sig-nificant difference between column and beam moment capacitytypical of the older practice in Italy as well as in other Mediter-ranean seismic-prone countries �where minimum values of longi-tudinal reinforcement ratio as low as 0.8% were allowed�.

Vertical FRP laminates were used on the external side �shear�face of the column in both interior and exterior joints �two layersper side� rather than on the flexural �tension and compression�side faces, in order to increase the column flexural capacity aswell as the joint shear strength. In addition, in the exterior jointspecimen, a U-shape horizontal laminate, wrapped around the ex-terior face of the specimen at the joint level, was used to increasethe joint shear strength as well as prevent the expulsion of aconcrete wedge. An adequately limited anchorage length withinthe beam was calculated in order to �a� guarantee sufficient shearstrengthening in the joint without excessively increasing the beamcapacity �as per Fig. 15� and �b� relocate the plastic hinge regionat a controlled distance from the beam-column critical interface.Although the evaluation of strengthening effects was carried outincluding debonding effects �when nonconservative�, additionalsmaller strips were used to wrap the main FRP laminates andprovide proper anchorage. In the case of the interior joint, theFRP laminate crossing the joint was intentionally left unprotectedfrom debonding in the joint panel zone region.

According to a partial retrofit approach, the retrofit solution forthe interior joint C3 �Fig. 14�, intended to allow some debondingof the vertical FRP sheets to occur along the joint panel zone, inorder to facilitate the development of a combined damage mecha-nism with limited cracking in the joint and subsequent flexuralhinging of the adjacent beams.

The target performance of the retrofit solution was controlledusing the proposed procedure based on the M-N performance-domain as shown in Fig. 15 and Table 3 for the exterior specimenT1B.

Prior to testing the beam-column specimens, a partial retrofitstrategy was implemented on the frame system �Fig. 16�, with thefinal intent to favor a more desirable inelastic global mechanism,able to protect brittle failure mechanisms due to the excessivedamage and collapse of an exterior joint or the development of asoft storey. This could be achieved by forming plastic hinges inthe exterior beams while accepting minor damage in the interiorjoint prior to the development of a flexural behavior in the adja-cent structural elements �in this case, the interior beams�. A simi-lar approach, in principle, and detailed layout of the FRP retrofitwas adopted as per the beam-column specimens. AppropriateM-N interaction curves, accounting for the effective geometryand demand curve for each beam-column joint within the framesystems, were used to verify the efficiency of the final solution,which, for simplicity of execution, was implemented for both thefirst and second floor joints, and followed the solution adopted forthe T1B and C3 specimens. No intervention seemed to be re-quired at the third-floor level where no or negligible damage wasexpected in the panel zone region with flexural cracks developingin the column top sections.

Fig. 18. Comparison of hysteresis response of as-built andFRP-retrofitted beam-column subassemblies: exterior joints T1 andT1B, T2 and T2B and interior joints C2 and C3

544

545546547548549550

551

552

553

554

555

556

557

558

559

560

561

562

563

564

565

566

567

568

569

570

571

572

573

574

575

576

577

578

579

580

581

582

583

584

585

586

587

588

589

590

591

592

593

594

595

596

597

598

599

600

601

602

603

604

605

606

607

608


Improvement of Structural Performanceafter Retrofit

Experimental Results on Retrofitted Beam-ColumnJoint Subassemblies

The results of the experimental quasistatic tests on three beamcolumn joints in the retrofitted configurations �namely T1B, T2B,and C3� provided very satisfactory confirmations of the efficiencyof the adopted retrofit solution as well as of the reliability ofthe analytical procedure developed to design the intervention

and assess the expected sequence of events and performance. Asummary of the results is given herein, while more detailsare available in Nassi �2002� and will be reported in futurepublications.

In all cases, the retrofit objective based on a multilevel retrofitstrategy was achieved, leading to a significant improvement in thebehavior of the subassemblies, which ultimately imply an en-hanced behavior of the frame system �adequate global inelasticmechanism�.

As shown in Fig. 17, a properly designed FRP-retrofit solutionfor exterior beam-column joints can protect and avoid the forma-tion of a brittle shear hinge mechanism and re-establish a moredesirable hierarchy of internal strengths and sequence of events,enforcing a beam plastic hinge mechanism, relocated at a con-trolled distance from the beam-column interface �total retrofit�.

As a result, an improved and more stable hysteresis behaviorwas observed with increased ductility and energy dissipation ca-pacity �Fig. 18�.

The values of lateral force corresponding to the occurrence ofthe critical events were well-predicted by the analytical methods�presented in Fig. 15 and Table 4�.

Similar considerations can be derived for the enhanced re-sponse of the interior joint specimen C3, where the partial retrofitstrategy led to a controlled debonding of the column vertical fi-bers crossing the joint �Fig. 7�. The formation of flexural damagein the column was thus postponed. In addition to the increasedoverall strength �as shown by the hysteresis loop in Fig. 18�,the FRP provided a favorable confinement effect in the column

Table 4. Sequence of Events for Exterior Specimen T1B �RetrofittedConfiguration�

Specimen T1B �strengthened�

Type oflateral force N° Event

Lateralforce�kN�

1 Beam yielding −18.91

Open joint F�0 2 Upper column yielding −23.11

3 Joint cracking�no strength degradation�

−24.15

4 Lower column yielding −25.32

5 Lower column yielding 15.75

Close joint F�0 6 Upper column yielding 16.98

7 Beam yielding 18.91

8 Joint failure 19.67

Fig. 19. Damage observations in the retrofitted test-frame �at 2% top drift�

609

610

611612

613614615616617

618

619620621622623624625626627628629630631632633634635636637638639640641642643644645


plastic hinge region avoiding the premature crushing and spallingof concrete cover, protecting from strength degradation, bucklingof the longitudinal bars, and consequent failure.

Experimental Results on Retrofitted Frame System

The global behavior of the retrofitted frame systems followed theexpectations and analytical predictions. As shown in Fig. 19,the frame response was characterized by the formation of plastichinges in the exterior beams, with no damage in the beam-columnjoints, protected by the FRP. In the interior joints, flexural verticalcracks developed at the beam-column interface at an earlierstage �0.5–0.6 % interstorey drift� and further extended within thejoint panel zone, confirming that cracking was occurring in thejoint. Due to the lower level of imposed drift, when compared tothe tests on subassemblies, the debonding of the FRP sheets alongthe joint region did not develop up to a complete peeling-offphenomenon.

As a consequence of the inverted hierarchy of strength, at leastin the exterior beams �partial retrofit�, a more desirable inelasticmechanism occurred, leading to higher strength and dissipationcapacity, as evident from the more stable global hysteresis loopshown in Fig. 9, up to higher level of drift �2%� when comparedto the as-built solution, before observing a softening behavior�onset of strength reduction� mainly due to P-D effects.

It is worth noting that the unloading global behavior of theretrofitted frame shows a loss of stiffness with some pinchingphenomenon similar to, although less evident than, that observedin the as-built system. As anticipated for the beam-column sub-assemblies, this effect can be due to the shear cracking develop-ing in both the exterior and the interior beam-column jointbelonging to the frame system, underneath the layers of FRP. Inline with the proposed analytical model, FRP and concrete con-tribute in parallel to the overall strength degradation curve of thejoints �Fig. 12�. An appropriate retrofit strategy would thus protectthe joint from excessive deformation �concentrated in the beam

plastic hinge�, while, due to the alteration of the hierarchy ofstrength, higher nominal shear �or principle tensile� stresses mightdevelop in the joint, part of which still has to be taken by theconcrete component. Furthermore, the reinforcement details ofthe exterior joints in the frame systems are in general more simi-lar to those of the T1 specimen �see Figs. 1 and 2� which showed�consistently with the predicted performance-based M-N domain�a more remarkable pinching behavior than the T2 specimen eitherbefore or after the retrofit intervention �Fig. 18�. In the case of theT2 specimen, in fact, lower beam reinforcement was adopted,leading to an earlier formation of a plastic hinge in the adjacentbeam, with less rotational demand, thus damage, in the panel zoneregion �see Figs. 20 and 21.

Concluding Remarks

The experimental results of quasistatic tests on beam-columnjoint specimens and three-storey frame systems, designed forgravity load only and retrofitted with CFRP laminates, providedvery satisfactory confirmation of the efficiency of similar solu-tions for existing buildings.

Fig. 20. Shear hinge damage mechanism in as-built exterior jointspecimen T1: formation of a concrete wedge mechanism

Fig. 21. Deformed shape of the retrofit frame during testing

646

647648

649

650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679

680

681682683684685686687688689690691692

693

694695696697698


A multilevel retrofit strategy has been proposed depending onthe subassembly type and structural details to achieve the desiredperformance with a feasible intervention. Alternative FRP com-posite materials in terms of mechanical properties or type �sheets,strips, or near mounted surface rods� should be appropriately se-lected depending on the target performance as well as on theoriginal “distance” between events according to hierarchy ofstrength considerations.

A simplified analytical procedure to evaluate and control thesequence of events using a M-N performance domain has beenpresented, after receiving promising confirmations from the satis-factory experimental results. In the exterior joints, the occurrenceof a brittle joint shear mechanism was adequately protected and amore desirable hierarchy of strengths and sequence of eventsachieved, leading to a more ductile and dissipating hysteresis be-havior. In the interior joints, a controlled minor cracking in thejoint panel zone was accepted, in order to protect a column swaymechanism.

At a global level, the implementation of a partial retrofitstrategy on a three-storey three-bay frame system favored thedevelopment of a more appropriate global inelastic mechanism,preventing brittle failure in exterior joints or undesired eventssuch as a soft storey mechanism.

Ultimately, as discussed in the Introduction, issues of accessi-bility of the joint region and invasiveness will have to be faced inreal applications. However, it is worth noting that a typical geo-metrical and plan configuration of existing buildings designed forgravity load only in the 1950s–1970s period consist of framesrunning in one direction only and lightly reinforced slab in theorthogonal direction, the latter being quite typical of the construc-tion practice in Mediterranean countries. In these cases, the adop-tion of the proposed retrofit intervention can be somehow facili-tated, when compared with more recently designed buildings withframes in both directions and cast-in-situ concrete slabs providingflange effects.

Acknowledgments

The financial support provided by the Italian Ministry of theUniversity and the University of Pavia, under a coordinatednational project �PRIN 2001�, as well as by the European Com-munity �Contract No. SPEAR G6RD-CT-2001-00525� is grate-fully acknowledged. The writers wish to thank the MAC S.p.a.Treviso for providing the materials and technical assistancefor the retrofit intervention. The assistance and cooperation,during different phases of the project, of postgraduate stu-dents Mr. A. Vecchietti and Mr. R. Nassi are also gratefullyacknowledged.

References

Antonopoulos, C. P., and Triantafillou, T. C. �2002�. “Analysis ofFRP-strengthened RC beam-column joints.” J. Compos. Constr. 6�1�,41–51.

Antonopoulos, C. P., and Triantafillou, T. C. �2003�. “Experimentalinvestigation on FRP-strengthened RC beam-column joints.” J. Com-pos. Constr. 7�1�, 39–43.

Aycardi, L. E., Mander, J. B., and Reinhorn, A. M. �1994�. “Seismicresistance of R. C. frame structures designed only for gravity loads:Experimental performance of subassemblages.” ACI Struct. J. 91�5�,552–563.

Beres, A., Pessiki, S., White, R., and Gergely, P. �1996�. “Implications of

experimental on the seismic behaviour of gravity load designed RCbeam-column connections.” Earthquake Spectra 12�2�, 185–198.

Bing, L., Yiming, W., and Tso-Chien, P. �2002�. “Seismic behavior ofnon-seismically detailed interior beam-wide column joints. Part I:Experimental results and observed behavior.” ACI Struct. J. 99�6�,��–��.

Calvi, G. M., Magenes, G., and Pampanin, S. �2002a�. “Relevanceof beam-column damage and collapse in RC frame assessment.”J. Earthquake Eng. ��1�, 75–100

Calvi, G. M., Magenes, G., and Pampanin, S. �2002b�. “Experimental teston a three storey R. C. frame designed for gravity only.” 12th ECEE,London, Paper No. 727.

Dodd, L. L., and Restrepo, J. I. �1995�. “Model for predicting cyclicbehavior of reinforcing steel.” J. Struct. Eng. 121�3�, 433–445.

Dolce, M., Cardone, D., and Marnetto, R. �2000�. “Implementation andtesting of passive control devices based on shape memory alloys.”Earthquake Eng. Struct. Dyn. 29�7�, 945–968.

FIB �Federation International du Beton� �2001�. “Externally bonded FRPreinforcement for RC structures.” FIB Bulletin No. 14, Lausanne.

FIB �Federation International du Beton� �2006�. “Retrofitting of concretestructures by externally bonded FRPs with emphasis on seismic ap-plications.” FIB Bulletin No. 35, Lausanne.

Gergely, J., Pantelides, C. P., Nuismer, R. J., and Reaveley, L. D. �1988�.“Bridge pier retrofit using fiber reinforced plastic composites.”J. Compos. Constr. 2�4�, 165–174.

Gergely, J., Pantelides, C. P., and Reaveley, L. D. �2000�. “Shearstrengthening of RC T-joints using CFRP composites.” J. Compos.Constr. 4�2�, 56–64.

Hakuto, S., Park, R., and Tanaka, H. �2000�. “Seismic load tests oninterior and exterior beam-column joints with substandard reinforcingdetails.” ACI Struct. J. 97�1�, 11–25.

Hertanto, � �2006�. “Seismic assessment of pre-1970s reinforcedconcrete beam column joint subassemblies” M.E. dissertation, Dept.of Civil Engineering, Univ. of Canterbury, Christchurch, NewZealand.

Holzenkämpfer, P. �1994�. “Ingenieurmodelle des verbundes geklebterbewehrung für betonbauteile.” Ph.D dissertation, TU Braunshwieg �inGerman�.

Mander, J. B., Priestley, M. J. N., and Park, R. �1988�. “Theoreticalstress-strain model for confined concrete.” J. Struct. Eng. 114�8�,1804–1826.

Nassi, R. �2002�. “Seismic retrofit strategy for underdesigned beam-column subassemblies using FRP.” Laurea thesis, Dept. of StructuralMechanics, Univ. of Pavia �in Italian�.

Neubauer, �, and Rostásy, � �1997�. “Design aspects of concretestructures strengthened with externally bonded CFRP-plates,concrete�composites.” 7th Int. Conf. Structural Faults and Repair,Vol. 2, 109–118.

Pampanin, S., Calvi, G. M., and Moratti, M. �2002�. “Seismic behaviourof R. C. beam-column joints designed for gravity loads.” 12th ECEE,London, Paper No. 726.

Pampanin, S., and Christopoulos, C. �2003�. “Non-invasive retrofit ofexisting RC frames designed for gravity loads only.” FIB2003 Symp.Concrete Structures in Seismic Regions, Federation International duBeton, Athens.

Pampanin, S., Magenes, G., and Carr, A. �2003�. “Modelling of shearhinge mechanism in poorly detailed RC beam-column joints.” Proc.FIB Symp. Concrete Structures in Seismic Regions, Federation Inter-national du Beton, Athens, paper No. 171.

Pampanin, S., Christopoulos, C., and Chen, T-H. �2006�. “Developmentand validation of a haunch metallic seismic retrofit solution for exist-ing under-designed RC frame buildings.” Earthquake Eng. Struct.Dyn. 35�14�, 1739–1766.

Pantazopoulou, S. J., and Bonacci, J. F. �1994�. “On earthquake-resistance reinforced concrete frame connections.” Can. J. Civ. Eng.21, 307–328.

Pantelides, C. P., Clyde, C., and Reaveley, L. D. �2000�. “Rehabilitation

699

700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733

734

735736737738739740741742743744

745

746747748749750751752753754755756

757

758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823


of R/C building joints with FRP composites.” 12WCEE Auckland,Paper No. 2306.

Park, R. �2002�. “A summary of results of simulated seismic load tests onreinforced concrete beam-column joints, beams and columns withsubstandard reinforcing details.” J. Earthquake Eng. 6�2�, 1–27.

Priestley, M. J. N. �1997�. “Displacement-based seismic assessment ofreinforced concrete buildings.” J. Earthquake Eng. 1�1�, 157–192.

Prota, A., Manfredi, G., Nanni, A., and Cosenza, E. �2002�. “Capacityassessment of GLD RC frames strengthened with FR.” 12th EuropeanConference on Earthquake Engineering, London, Paper No. 241.

Regio Decreto. �1939�. 16 November-XVIII, n. 2228 �in Italian�.Santarella, L. �1957�. Il cemento armato, Vol. 2, Ed. U. Hoepli, Milan.

Spoelstra, M. R., and Monti, G. �1999�. “FRP-confined concrete model.”J. Compos. Constr. 3�3�, 143–150.

Sugano, S. �1996� “State of the art in techniques for rehabilitation ofbuildings.” Proc., 11th World Conf. on Earthquake Engineering,Acapulco, Mexico, Paper No. 2175.

Tsonos, A. D., and Stylianidis, K. A. �1999�. “Pre-seismic and post-seismic strengthening of reinforced concrete structural subassem-blages using composite materials �FRP�.” Proc., 13th Hellenic Con-crete Conference, Rethymno, Greece, 1, 455–466 �in Greek�.

Vecchietti, A. �2001�. “Seismic rehabilitation of concrete frame systemsdesigned for gravity loads only using FRP composite materials.” Lau-rea thesis, Dept. of Structural Mechanics, Univ. of Pavia �in Italian�.

824

825826827828829830831832833834835

836

837838839840841842843844845846847


performance-based seismic retroﬁt strategy for existing

Documents